From: Bob Johnson (robertjo_at_physics.upenn.edu)
Date: Thu May 10 2007 - 16:12:09 CDT

Hello everyone,
I have performed a replica exchange MD simulation of single stranded DNA (ssDNA)
adsorbed to a carbon nanotube (NT). I now want to analyze the trajectory and
categorize each ssDNA conformation by computing RMSDs with various reference
structures. However, using the normal 'measure fit' tool is going to yield
incorrect results.

The problem is, the ssDNA is adsorbed to a cylindrical surface. Thus, the
correct fit in my case is a RMSD fit under the restraint that the ssDNA remains
on the surface of the NT.

I've thought of a variety of ways to go about performing the alignment. First
would be to perform a translation in the z-direction (this is the axial
direction of the NT) followed by a rotation about the NT axis such that the
reference structure and the instantaenous ssDNA conformation's center of mass
would be aligned. However, I would also have to perform another rotation (about
the x-axis perhaps) to maximize the overlap of the atoms of each structure (make
sure that both structures are "pointing" in the same direction).

I wanted to see if anyone had some ideas about how this could be done. Also,
what kind of algorithm does the 'measure fit' tool use? Perhaps, I could write
a script/program that performs this fitting with the constraints I described
above?
Thanks,
Bob Johnson